# rmatrixt: Matrix t sampler In matrixsampling: Simulations of Matrix Variate Distributions

## Description

Samples the matrix t-distribution.

## Usage

 `1` ```rmatrixt(n, nu, M, U, V, checkSymmetry = TRUE, keep = TRUE) ```

## Arguments

 `n` sample size, a positive integer `nu` degrees of freedom, a positive number `M` mean matrix, without constraints `U` columns covariance matrix, a positive semidefinite matrix of order equal to `nrow(M)` `V` rows covariance matrix, a positive semidefinite matrix of order equal to `ncol(M)` `checkSymmetry` logical, whether to check the symmetry of `U` and `V` `keep` logical, whether to return an array with class keep

## Value

A numeric three-dimensional array; simulations are stacked along the third dimension (see example).

## Note

When `p=1` and `V=nu` or when `m=1` and `U=nu`, the distribution is the multivariate t-distribution.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16``` ```nu <- 4 m <- 2 p <- 3 M <- matrix(1, m, p) U <- toeplitz(m:1) V <- toeplitz(p:1) Tsims <- rmatrixt(10000, nu, M, U, V) dim(Tsims) # 2 3 10000 apply(Tsims, 1:2, mean) # approximates M vecTsims <- t(apply(Tsims, 3, function(X) c(t(X)))) round(cov(vecTsims), 1) # approximates 1/(nu-2) * kronecker(U,V) ## the `keep` class is nice when m=1 or p=1: Tsims <- rmatrixt(2, nu, M=1:3, U=diag(3), V=1) Tsims[,,1] # dimensions 3 1 # without `keep`, dimensions are lost: rmatrixt(2, nu, M=1:3, U=diag(3), V=1, keep=FALSE)[,,1] ```

matrixsampling documentation built on Aug. 25, 2019, 1:03 a.m.